Committee for Risk Assessment RAC Opinion Proposing

Committee for Risk Assessment RAC

Opinion proposing harmonised classification and labelling at EU level of

imidacloprid (ISO); (E)-1-(6-chloropyridin-3-ylmethyl)-N- nitroimidazolidin-2-ylidenamine

EC Number: 428-040-8 CAS Number: 138261-41-3

CLH-O-0000001412-86-282/F

Adopted 13 June 2019

Annankatu 18, P.O. Box 400, FI-00121 Helsinki, Finland | Tel. +358 9 686180 | Fax +358 9 68618210 | echa.europa.eu

[04.01-ML-014.03]

13 June 2019

CLH-O-0000001412-86-282/F

OPINION OF THE COMMITTEE FOR RISK ASSESSMENT ON A DOSSIER PROPOSING HARMONISED CLASSIFICATION AND LABELLING AT EU LEVEL

In accordance with Article 37 (4) of Regulation (EC) No 1272/2008, the Classification, Labelling and Packaging (CLP) Regulation, the Committee for Risk Assessment (RAC) has adopted an opinion on the proposal for harmonised classification and labelling (CLH) of:

Chemical name: imidacloprid (ISO); (E)-1-(6-chloropyridin-3-ylmethyl)-N- nitroimidazolidin-2-ylidenamine

EC Number: 428-040-8

CAS Number: 138261-41-3

The proposal was submitted by Germany and received by RAC on 3 September 2018.

In this opinion, all classification and labelling elements are given in accordance with the CLP Regulation.

PROCESS FOR ADOPTION OF THE OPINION Germany has submitted a CLH dossier containing a proposal together with the justification and background information documented in a CLH report. The CLH report was made publicly available in accordance with the requirements of the CLP Regulation at http://echa.europa.eu/harmonised-classification-and-labelling-consultation/ on 8 October 2018. Concerned parties and Member State Competent Authorities (MSCA) were invited to submit comments and contributions by 7 December 2018.

ADOPTION OF THE OPINION OF RAC Rapporteur, appointed by RAC: Annemarie Losert

Co-Rapporteur, appointed by RAC: Pietro Paris

The opinion takes into account the comments provided by MSCAs and concerned parties in accordance with Article 37(4) of the CLP Regulation and the comments received are compiled in Annex 2. The RAC opinion on the proposed harmonised classification and labelling was adopted on 13 June 2019 by consensus.

Classification and labelling in accordance with the CLP Regulation (Regulation (EC) 1272/2008)

Index No Chemical name EC No CAS No Classification Labelling Specific Conc. Notes Hazard Class and Hazard Pictogram, Hazard Suppl. Limits, M- Category Code(s) statement Signal Word statement Hazard factors and Code(s) Code(s) Code(s) statement ATE Code(s) Current 612-252- imidacloprid (ISO); - 138261- Acute Tox. 4 * H302 GHS07 H302 Annex VI 00-4 (E)-1-(6- 41-3 Aquatic Acute 1 H400 GHS09 H410 entry chloropyridin-3- Aquatic Chronic 1 H410 Wng ylmethyl)-N- nitroimidazolidin-2- ylidenamine Dossier 612-252- imidacloprid (ISO); - 138261- Modify Modify Modify Modify Add submitters 00-4 (E)-1-(6- 41-3 Acute Tox. 3 H301 GHS06 H301 Oral: ATE = 131 proposal chloropyridin-3- Dgr mg/kg bw ylmethyl)-N- Retain Retain Retain nitroimidazolidin-2- Aquatic Acute 1 H400 Retain H410 M=100 ylidenamine Aquatic Chronic 1 H410 GHS09 M=1000 RAC opinion 612-252- imidacloprid (ISO); - 138261- Acute Tox. 3 H301 GHS06 H301 Oral: ATE = 131 00-4 (E)-1-(6- 41-3 Aquatic Acute 1 H400 GHS09 H410 mg/kg bw chloropyridin-3- Aquatic Chronic 1 H410 Dgr M=100 ylmethyl)-N- M=1000 nitroimidazolidin-2- ylidenamine Resulting 612-252- imidacloprid (ISO); - 138261- Acute Tox. 3 H301 GHS06 H301 Oral: ATE = 131 Annex VI 00-4 (E)-1-(6- 41-3 Aquatic Acute 1 H400 GHS09 H410 mg/kg bw entry if chloropyridin-3- Aquatic Chronic 1 H410 Dgr M=100 agreed by ylmethyl)-N- M=1000 COM nitroimidazolidin-2- ylidenamine

GROUNDS FOR ADOPTION OF THE OPINION

RAC general comment

Imidacloprid is an active ingredient in biocidal and plant protection products. Biocidal products containing imidacloprid are intended for professional use (e.g. by pest control operators, farmers), in bait formulations controlling insects such as house flies and cockroaches. The pesticidal product is currently restricted for use as an insecticide to green houses only. Imidacloprid belongs to the family of neonicotinoids and has an existing harmonised classification and labelling in Annex VI to CLP, which was introduced with the first ATP by translation from a previous harmonised classification.

HUMAN HEALTH HAZARD EVALUATION

RAC evaluation of acute toxicity

Summary of the Dossier Submitter’s proposal

Animal data The data on acute oral toxicity presented in the CLH dossier consisted of five oral acute toxicity studies, all according to OECD TG and GLP. There are three rat studies and one mouse study according to OECD TG 401, which were all conducted at the same laboratory and all four used Cremophor EL (2% v/v water) as a vehicle. An additional rat study according to OECD TG 423 was conducted at a different laboratory, using 0,5% aqueous carboxymethyl cellulose as vehicle. Three different batches of imidacloprid were tested but the same batch was used in the mouse study and in two rat studies. For this repeatedly used batch a slightly different purity (94.2% vs 94.3%) has been determined in independent analyses.

Table: Animal data

Guideline Species, Dose levels Results Remarks References Strain Frequency of Route, Sex application Species No of GLP animals OECD TG Wistar rat, Dose levels: 50 in LD50: clinical signs Anon 1, 401 males & 100, 250, 424 from 100 mg/kg 1989a 5/sex in each 315, 400, 450, mg/kg GLP; bw in males & dosage 500, 1800 mg/kg bw 250 mg/kg bw in oral/ group, bw in males & (males) females gavage; females; in & 450- onwards; Vehicle: rat; females 475 Cremophor mixed additionally 475 mg/kg first dose EL (2% v/v batch mg/kg bw were bw causing water) 180587, applied; (females mortality: 400 purity ) mg/kg bw in 94,2% Single application males & females

OECD TG Wistar rat, Dose levels: 50, LD50: clinical signs Anon 2, 401 200, 350, 400, 642 from 200 mg/kg 1991a 5/sex in each 500, 600, 750, mg/kg GLP; bw in males & dosage 1000 mg/kg bw in bw 400 mg/kg bw in oral/ group, males & 100, 400, (males) females gavage; 450, 500, 600, and 648 onwards; Vehicle: rat; 1000 mg/kg bw in mg/kg Cremophor batch females; bw first dose EL (2% v/v 17133/90, (females causing water) purity 96% Single application ) mortality: 350 mg/kg bw in males & 450 mg/kg bw in females

OECD TG Wistar rat, Dose levels: 50, LD50: clinical signs Anon 3, 401 200, 300, 350, 504 from 200 mg/kg 1991b 5/sex in each 400, 500, 600 mg/kg GLP; bw onwards in dosage mg/kg bw in bw males & females; oral/ group, males & 100, 200, (males) gavage; 300, 350, 400, and 379 first dose Vehicle: rat; 500 mg/kg bw in mg/kg causing Cremophor mixed females; bw mortality: 300 EL (2% v/v batch (females mg/kg bw in water) 180587, Single application ) males & females purity

94,3%

Guideline Species, Dose levels Results Remarks References Strain Frequency of Route, Sex application Species No of GLP animals OECD TG Wistar rat, Dose levels: 300 & LD50 At the limit dose Anon 4, 423 2000 mg/kg bw; between of 2000 mg/kg 2006 3 females/ 300 & GLP; bw all animals dosage Single application 2000 died on the day oral/ group, mg/kg of dosing. gavage; bw rat; One group at No unscheduled 2000 mg/kg batch SI- mortality in rats bw, two 06016, at 300 mg/kg groups at purity bw. 300 mg/kg 98,56% bw Lethargy and tremor preceded Vehicle: mortality at the 0.5% high dose, no aqueous signs were seen carboxymeth at the low dose. yl cellulose

OECD TG Bor:NMRI Dose levels: 10, LD50: clinical signs Anon 5, 401 mice, 71, 100, 120, 140, 131 from 71 mg/kg 1989b mg/kg GLP; 160, 250 mg/kg bw in males & 5/sex in each bw in males & 10, bw 100 mg/kg bw in oral/ dosage 100, 120, 140, (males) females; gavage; group, 160, 250 mg/kg and 168 mouse; bw in females; mg/kg first dose Vehicle: mixed bw causing Cremophor batch Single application (females mortality: 100 EL (2% v/v 180587, ) mg/kg bw in water) purity males & 120 94,2% mg/kg bw in females.

While all rat studies support a classification as Acute Tox. 4, the single mouse study supports a classification as Acute Tox. 3. The dossier submitter (DS) concluded that the mouse was more sensitive than the rat and that male rats were more sensitive than female rats, based on the observed clinical findings and LD50 values in males and females. Clinical signs consisted among others of apathy, laboured breathing, accelerated breathing, decreased motility, staggering gait, narrowed eyelids, trembling and spasms, transient tremor and convulsions, transient or continuing spasms, salivation, increased water intake, diuresis, piloerection and absence of faeces. Some of these effects are indicative of neurotoxicity. At necropsy, no test substance-related changes were noted in surviving animals which were killed at the end of the observation period. No studies for dermal and inhalation route are available.

Human data

As for most pesticides, no information is available on effects in humans due to exposure to the active ingredient, imidacloprid, itself. The DS stated that according to the DAR (2006), occupational medical surveillance of employees in manufacturing the substance did not reveal indications of adverse effects but there are no more recent data. However, the CLH report lists a number of published clinical and forensic case reports on poisoning incidents with various plant protection products containing imidacloprid. The DS also mentions a report by Proença et al. (2005), which has been available for the evaluation of imidacloprid as a biocide already, but had not been taken into consideration by EFSA or ECHA so far. The data consist of eight poisoning cases, four of which were lethal and four in which the patients survived (see tables in the Background Document). The actual intake of imidacloprid in these cases is not precisely known, but it can be roughly estimated, at least for some cases.

Overall, the DS concluded that cardiac toxicity seems to be of particular importance and critical for the outcome of the reported human poisoning cases, neurotoxicity seems less important. The dossier submitter further concluded that the toxicity observed in these cases appears to be considerably stronger than in experimental animals and traces this back to irritant/corrosive solvents (like e.g. N-methyl pyrrolidone, NMP) present in the formulations, which were ingested or inhaled in the human poisoning cases.

The dossier submitter was of the view that the constituents contained in the formulation are responsible for the more severe toxicity seen in humans, rather than assuming a higher sensitivity of humans towards imidacloprid. The difference between humans and experimental animal species would be so big that this would be hardly conceivable.

The DS concluded that imidacloprid was moderately toxic to rats whereas mice proved more sensitive. Human experience points to higher toxicity of formulations. The DS mentioned that the irritant/corrosive properties of solvents such as N-methyl pyrrolidone (NMP) could be responsible for the higher toxicity of formulations. They emphasised that only the oral route was considered in detail for classification.

Based on the lowest LD50 of 131 mg/kg bw, as determined in male mice, the DS proposed to classify imidacloprid as Acute Tox. 3; H302, with an ATE of 131 mg/kg bw. They also argued that according to the CLP Regulation, the classification should be based on the lowest determined

LD50, if coming from a reliable study. The mouse study was carried out at the same laboratory, using the same vehicle as in all three rat studies and using the same batch as in two of the rat studies, which further supports that mice have a higher sensitivity towards imidacloprid than rat.

The LD50 of 131 mg/kg bw in males is close to the LD50 of 168 mg/kg bw in female mice, further supporting that the mouse is more sensitive than the rat. There is no information available that the mouse would be less relevant for the assessment of imidacloprid’s acute toxicity. The dossier submitter concluded that the human poisoning cases were supportive only, as they all resulted from imidacloprid up-take of formulations, which contain solvents, which are likely to have increased the toxicity of the formulation compared to pure imidacloprid.

Comments received during public consultation

During the public consultation one comment was received from the Manufacturer in favour of keeping the classification in Category 4 including the * for minimum classification. In their comment, the Manufacturer described that based on the studies available at that time imidacloprid was classified as R22 (harmful if swallowed) included in the 31st adaptation to technical progress (CD 2009/2/EC). This classification was later translated into a classification

according the CLP Regulation as Acute Tox. 4*. The Manufacturer stated that an asterisk was added to mark it as a minimum classification in view of the available data from the mice study, indicating higher toxicity. They also stated that the DS based their proposal to change from Acute Tox. Category 4* to Category 3 on (a) that there is no information in the guidance on species relevance that would allow disregarding the finding in mice and (b) additional evidence coming from poisoning incidents in humans that acute oral toxicity of imidacloprid might be of concern.

The manufacturer was of the view that the current minimum classification would be in place to cover the existence of data which show higher toxicity (i.e. the mouse data would not be disregarded as recommended by the guidance).

The manufacturer also stated that the DS’ proposal to change classification to Acute Tox. 3 was based on human poisoning cases. They listed several drawbacks of the human data including too low intake amounts, exposure to formulations with effects from other constituents, lacking information on intake amounts and other relevant information and that for one case no signs of intoxication were evident. The DS responded that the current classification proposal is based on the results of a valid oral acute toxicity study in the mouse, which demonstrates that this species is more sensitive than the rat. In line with the manufacturer, the dossier submitter was of the opinion that the current guidance document does not say that a certain species would not be relevant. The dossier submitter also supported the Manufacturers view that the human poisoning incidents should not have an impact on the classification in this case. RAC agrees with the DS’ response.

Assessment and comparison with the classification criteria

RAC concludes that the available animal studies clearly indicate that imidacloprid is acutely toxic via the oral route. While the rat data would support a classification in Category 4, the single mouse study supports a classification in Category 3. As all 5 studies were conducted according to guideline and GLP and had no drawbacks, no difference regarding their suitability for the assessment of classification can be made. As the study in the mouse is no less relevant for humans, the classification has to be based on this, as the most sensitive species, i.e. in line with chapter 3.1.2.3.2 of the “Guidance on the Application of the CLP Criteria” (Version 5.0, July 2017.

RAC also reviewed the human poisoning cases and agrees with the DS and the Manufacturer that they should not have an impact on classification. One major drawback of the available human data is that the exposure was to insecticide formulations, not to pure imidacloprid and that exposure levels could only be roughly estimated. It is further noted that cardiac toxicity was of particular importance in those cases, rather than neurotoxicity and on balance the human cases suggest a markedly higher toxicity of formulations as compared to the active substance. While it might be possible that there are differences regarding the toxicological profile between animals and human, it is rather unlikely that the symptoms or the fatal outcomes may be attributed to imidacloprid alone. If so, acute toxicity in humans would be considerably higher than in animals. Constituents of these mixtures, like solvents with irritant / corrosive or other toxic properties have to be considered and are likely to have contributed to the toxicity of the mixtures. In conclusion, RAC is of the view that the human poisoning cases indicate that acute toxicity of such formulations is of concern, but they cannot be used to decide on the classification of pure imidacloprid.

Comparison with the criteria

Based on the lowest LD50 value of 131 mg/kg bw derived for male mice in an acute toxicity study in mouse (Anon 5, 1989b), RAC supports a classification as Acute Tox. 3; H301 (LD50 > 50 but ≤ 300mg/kg bw) in line with the DS proposal.

RAC also supports to use the same LD50 values of 131 mg/kg bw as ATE for imidacloprid.

ENVIRONMENTAL HAZARD EVALUATION

RAC evaluation of aquatic hazards (acute and chronic)

Summary of the Dossier Submitter’s proposal

Imidacloprid is an insecticide for plant protection and biocidal products. An environmental harmonised classification as Aquatic Acute 1; H400 and Aquatic Chronic 1; H410, with no M- factors can be found in Annex VI of the CLP, which was translated from the classification under the previous legislation (DSD 67/548/EEC). The DS proposal for environmental classification was based on new information on aquatic toxicity, which confirms the existing hazard categories and adds appropriate M-factors. The key studies are non-guideline using various non-standard freshwater invertebrate species.

Degradation In a hydrolysis study (Yoshida, 1989) conducted according to US EPA Guideline § 161-1 and in compliance with GLP, imidacloprid was incubated at 25 °C for 30 days in pH 5, 7 and 9 aqueous solutions. Imidacloprid was found to be stable at pH 5 and 7. Slow hydrolysis with a half-life of approximately 1 year occurred at pH 9 (DT50 = 2.75 years, calculated by the DS at 12.5 °C). No significant hydrolysis products were determined. The photodegradation of radio-labelled imidacloprid in water was studied according to three guidelines US EPA § 161-2, “Photo-transformation of Chemicals in Water”, UBA, Germany, Nov. 1990, and OECD TG 316. The studies showed that imidacloprid was rapidly photodegraded in water with half-lives < 1 day. Photodegradation involved the formation of up to 15 phototransformation products, among them four that reached levels higher than 10% of the applied radioactivity.

No ready biodegradability studies were performed.

An aerobic mineralisation in surface water study (Stevens et. al., 1997), comparable to OECD TG 309, showed that imidacloprid disappears slowly in non-sterile, non-light exposed test system with a DT50 of 331 days at 22°C and a mineralisation rate of 4.3% after 366 d. Imidacloprid was metabolised into nine quantifiable degradation products, among them NTN33893-desnitro, which exceeded 10% of the initially applied radioactivity. No information is available as to whether the degradation products are hazardous to the aquatic environment. The aerobic transformation of radiolabelled imidacloprid was investigated in two water/sediment studies, conducted according to US EPA § 162-4 and in compliance with GLP. In the first study (Wilmes, 1990), the dissipation behaviour of imidacloprid applied at a concentration of 0.2 mg/L for a 10 cm deep water body was studied in two Dutch water-sediment systems in the dark at 22 ± 1 °C over a period of 92 days. The half-lives (whole system) for the dissipation of imidacloprid calculated according to first-order kinetics were found to be 32 and 142 days for the two systems. CO2 was formed in both test systems in small quantities (1.4 and 2.0% of the

applied radioactivity). Three metabolites were detected in the water phase and the sediment in both test systems. In one system, none of them reached a level of >10% of AR. In the other one, one metabolite (NTN33893-desnitro) reached a level of 12.3% (sum of amounts found in water and sediment) at the end of the study.

A third water-sediment system, originating from the USA (Stilwell, Kansas) was investigated under aerobic conditions in the dark at 22 ± 1 °C over a period of 30 days (Spiteller, 1993). A first order half-life of 129 days was calculated at 20 °C for the whole system. Negligible mineralisation occurred since 0.7% of the applied radioactivity had been completely mineralised at the end of study. Four metabolites were identified as minor metabolites but none reached a level of 10% of applied radioactivity. Aerobic degradation in soil was investigated in five laboratory studies with European soils at 20 °C in the dark. First-order half-lives varied between 106 days and 193 days. Mineralisation was limited, accounting for a maximum of 20.3% in one sandy loam soil after 126 days. In total, nine different degradation products have been identified, none exceeding 5%. In conclusion, the DS considered imidacloprid to be not rapidly degradable as it is hydrolytically stable and not ultimately degraded to a level greater than 70% over 28 days in surface water, water/sediment and soil simulation studies.

Bioaccumulation

Based on experimental data, imidacloprid has a measured log KOW of 0.57 (OECD TG 107, 21 °C and pH 7). A study on the bioaccumulation behaviour of imidacloprid is not available. The BCF for fish has been predicted from the linear relationship between KOW and BCF developed by Veith et al. (1979). According to the “Technical Guidance Document on Risk Assessment Part III”, relevant under the Biocidal Products Directive (BPD 98/8/EG), the linear model generated by Veith et al., (1979)

(log BCFfish = 0.85log KOW - 0.70) can be used for substances with a log KOW < 6. Therefore the calculated value of the bioconcentration factor for imidacloprid in fish on a wet weight basis is

BCFfish = 0.609 L.Kgwetfish.

Based on a measured log kOW of 0.57 being below the CLP criterion of 4, the DS considered imidacloprid to have a low potential for bioaccumulation.

Aquatic Toxicity

Studies on acute and long-term aquatic toxicity to imidacloprid for all three trophic levels are available. Studies are also available for the main metabolite (NTN33893-desnitro).

The test results are summarised in the following table. The key tests forming the basis for classification are reported in bold.

Table: Summary of information most relevant for classification on aquatic toxicity

Results

Test Test LC50/EC5 Test Method NOEC Reference organism system Endpoint 0 concentration [mg/L] [mg/L]

Fish

Nominal Oncorhynch Static (confirmed by Anonymous OECD TG 203 Mortality 211 us mykiss 96h analytical , 1988 monitoring)

Results

Test Test LC50/EC5 Test Method NOEC Reference organism system Endpoint 0 concentration [mg/L] [mg/L] U.S.-EPA- FIFRA, 40 Anonymous Oncorhynch Static Mean CFR, Section Mortality > 83 , 1990 us mykiss 96h measured 158.145, Guideline 72-1 EEC Nominal Anonymous DIRECTIVE Leuciscus Static (confirmed by Mortality 237 , 1987 79/831/WG, idus 96h analytical Annex V monitoring)

Anonymous Oncorhynch Flow- Time to hatch Mean OECD TG 210 9.02 , 2002 us mykiss through and swim up measured 91d

Aquatic invertebrates

Nominal Non concentration Roessink Cloeon Static, Immobilisatio guideline 0.00102 s (confirmed et al., dipterum 96h n study by analytical 2013 monitoring) Same experimental setup and species as Roessink et al. Van den Non guideline Cloeon Static, 2013, but Immobilisation 0.018 Brink et al., study dipterum 96h instead of 2016 summer generations, winter generations were tested Nominal Non guideline Caenis Static, (confirmed by Roessink et Immobilisation 0.00177 study horaria 96h analytical al., 2013 monitoring) Same experimental setup and species as Roessink et al. Van den Non guideline Caenis Static, 2013, but Immobilisation 0.0060 Brink et al., study horaria 96h instead of 2016 summer generations, winter generations were tested.

Nominal Plea Non guideline Static, (confirmed by Roessink et minutissim Immobilisation 0.0359 study 96h analytical al., 2013 a monitoring)

Daphnia Static, Mean Young and OECD TG 202 Immobilisation 85 magna 48h measured Hicks, 1990

Results

Test Test LC50/EC5 Test Method NOEC Reference organism system Endpoint 0 concentration [mg/L] [mg/L] Nominal Notonecta (confirmed by Roessink et Non guideline Static, 0.0182 spp. Immobilisation analytical al., 2013 study 96h monitoring)

Nominal Non guideline Limnephilid Static, (confirmed by Roessink et Immobilisation 0.00179 study ae 96h analytical al., 2013 monitoring) Nominal Non guideline Asellus Static, (confirmed by Roessink et Immobilisation 0.119 study aquaticus 96h analytical al., 2013 monitoring) Nominal Non guideline Chaoborus Static, (confirmed by Roessink et Immobilisation 0.284 study obscuripes 96h analytical al., 2013 monitoring) Nominal Non guideline Sialis Static, (confirmed by Roessink et Immobilisation 0.0506 study lutaria 96h analytical al., 2013 monitoring) Nominal Dorgerloh Chironomus Static, (confirmed by and OECD TG 202 Mortality 0.055 riparius 24h analytical Sommer, monitoring) 2002a US EPA FIFRA, England 40 CFR, Part Hyalella Static, Mortality 0.526 Mean and 158.145 azteca 96h measured Bucksath, Guideline No. immobilty 0.055 1991 72-2 US EPA FIFRA, 40 CFR, Part Flow- Mysidopsis Mean 158.145 through Mortality 0.034 Ward, 1990 bahia measured Guideline No. , 96h 72-2 Nominal Semi- Non guideline Cloeon 0.000033 (confirmed by Roessink et static Immobilisation study dipterum (EC10) analytical al., 2013 28d monitoring) Nominal Non Semi- Roessink Caenis Immobilisatio 0.00002 (confirmed by guideline static et al., horaria n 4 (EC10) analytical study 28d 2013 monitoring)

Same experimental setup and species as Roessink et al. Semi- Van den Non guideline Cloeon 2013, but static Immobilisation 0.4 (EC10) Brink et al. study dipterum instead of 28d 2016 summer generations, winter generations were tested

Nominal Semi- Non guideline Asellus (confirmed by Roessink et static Immobilisation 0.00171 study aquaticus analytical al., 2013 28d monitoring)

Results

Test Test LC50/EC5 Test Method NOEC Reference organism system Endpoint 0 concentration [mg/L] [mg/L] Nominal Semi- Non guideline Gammarus (confirmed by Roessink et static Immobilisation 0.00295 study pulex analytical al., 2013 28d monitoring)

Nominal Semi- Non guideline Chaoborus (confirmed by Roessink et static Immobilisation 0.00457 study obscuripes analytical al., 2013 28d monitoring)

Nominal Semi- Non guideline Sialis (confirmed by Roessink et static Immobilisation 0.00128 study lutaria analytical al., 2013 28d monitoring)

Nominal Plea Semi- 0.00203 Non guideline (confirmed by Roessink et minutissim static Immobilisation study analytical al., 2013 a 28d 0.00645 monitoring)

Semi- Young and US EPA-FIFRA Daphnia Reproduction, 1.8 Mean static Blakemore, 72-4 magna survival, length (length) measured 21d 1990

OECD TG 202

Test 51.8 substance: (LC50) Mean Roney and Hyalella Static Mortality measured Bowers, azteca 96h Major 29.8 concentration 1996 metabolite (EC50) imidacloprid desnitro

Algae and aquatic plants

Scenedesm Nominal Static Heimbach, OECD TG 201 us Growth rate >10 ≥10 96h Limit test with 1986a subspicatus 10 mg/L Nominal Selenastru (Confirmed by m Static analytical Dorgerloh, OECD TG 201 Growth rate >100 <100 capricornut 72h monitoring) 2000 um Limit test with 100 mg/L

Other aquatic organisms (including sediment)

Nominal Dorgerloh Chironomus Static Development, 0.00209 and OECD TG 219 0.00311 Mean riparius 28d Emergence 0.00087 Sommer, measured 2001a Based on guidelines by Semi- ASTM (1988, Chironomus Growth, Mean Gagliano, static 0.00317 0.00067 1990) and tentans survival measured 1991 10d USEPA (1975, 1982, 1985) OECD TG 219 Nominal Dorgerloh Major Chironomus Static Development, 0.046 0.027 and metabolite Mean riparius 28d Emergence 0.00945 Sommer, imidacloprid measured 2001b desnitro

Acute toxicity

Three acute toxicity studies to fish are available and included in the CLH Report. In the reliable study by Anonymous (1988), the short-term toxicity of imidacloprid (technical active substance) was examined on young rainbow trout (Oncorhynchus mykiss) under static condition and according to OECD TG 203. A 96h LC50 of 211 mg/L based on nominal concentrations (analytically confirmed - measured > 80% of nominal concentrations) was determined. This study is considered as acceptable with fulfilled validity criteria and used as relevant data for purpose of the acute classification. Further short term fish toxicity studies Anonymous (1990) and Anonymous (1987), conducted respectively with Oncorhynchus mykiss and Leuciscus idus according to EPA Guideline 72-1 and EEC directive 79/831/WG-1984, are reported as adequate acute toxicity data and used as supplementary information.

Based on the available data, imidacloprid shows a low acute toxicity to fish with a reliable LC50 value >1 mg/L. Short-term toxicity tests with 10 aquatic invertebrate species from different taxonomic groups are available (Roessink et al. 2013). Test organisms were collected from an uncontaminated aquatic ecosystems. Early larval insect instars were used for the tests. The test organisms were acclimated for at least 3 days to laboratory conditions (18 +/- 2 °C, 12:12 hours light: dark). The exposure period was 96h and the endpoints used were immobilisation and mortality. Imidacloprid concentrations measured in the dosing solution were, on average, 97.5% of the nominal concentration. No further analytical monitoring was performed. However, from the analytical monitoring performed for the long-term studies the DS concluded that the test substance concentration was stable during the exposure period of 96h and thus the use of nominal concentrations is justified. Concerning the validity criteria of OECD TG 202 (Daphnia Acute Immobilisation Test), the criterion of 10% maximum immobilisation in controls is fulfilled for 8 of the 10 tests. In the summarising table above, only studies with reliability Klimisch score

1 and 2 were reported. The 96h EC50 values range from 1.02 – 284 μg/L for the endpoint immobilisation. The most sensitive species were Cloeon dipterum (1.02 μg/L), Caenis horaria (1.77 μg/L) and Limnephilidae (1.79 μg/L). The DS also provided another non guideline study (van den Brink et al. 2016) with the same experimental setup and Ephemeroptera species as Roessink et al. (2013) but instead of summer generations, winter generations were tested. The short-term toxicity values for the winter generations are higher than for the summer generations. Therefore, the DS proposed to use for classification the lowest toxicity values for the summer generations, as most relevant for hazard assessment.

No effects were seen in two limit tests with green algae at the concentration of 10 mg/L and 100 mg/L.

Based on the 96h EC50 (immobilisation) of 0.00102 mg/L for Cloeon dipterum, the DS proposed classification as Aquatic Acute 1 (M=100).

Chronic toxicity A single chronic toxicity study on fish performed with Imidacloprid is provided in the CLH Report. In this study (Anonymous, 2002), the long term toxicity of Imidacloprid (technical active substance) was tested on Oncorhynchus mykiss in a fish early life-stage study conducted according to OECD TG 210. Observed endpoints were time to hatch and hatching rate, larval deformities and survival, time to swim-up, behavioural changes and post-hatch survival and growth. Based on mean measured concentrations, the NOEC was determined to be 9.02 mg/L for the most sensitive endpoints (time to hatch and swim up). For the other observed endpoints

the NOEC was 26.9 mg/L. This study is considered valid and useful for purpose of chronic classification.

Based on the available data, Imidacloprid shows a low chronic toxicity to fish, with a lowest NOEC of 9.02 mg/L.

Regarding aquatic invertebrates, long-term toxicity tests with species from 7 different taxonomic groups were available (Roessink et al., 2013). Test organisms were collected from an uncontaminated aquatic ecosystem. Early larval insect instars were used for the tests, after acclimation for at least 3 days to laboratory conditions (18 ± 2 °C, 12:12 hours light: dark). Five concentrations and a control were tested using 3 replicates with each 10 test animals. Immobilisation and mortality were the detected endpoints for an exposure period of 28 d. Every week the test solution was renewed. Imidacloprid concentrations measured in the dosing solution were, on average, 95.5% of the nominal concentration. Analytical monitoring was performed for the control and the highest test concentration. Measured concentrations were in the range of 84.9 – 97% of the nominal concentration, thus proving the test substance to be stable during the exposure phase. 28d EC10 values (immobilisation) for the 7 tested species were in the range of 0.024 – 4.57 μg/L.

As in the short-term studies, the mayflies Cloeon dipterum (28d EC10 = 0.033 μg/L) and Caenis horaria (28d EC10 = 0.024 μg/L) were most sensitive. The DS provided also the result from non-guideline study by Van den Brink et al. (2016) with Cloeon dipterum. The experimental setup was the same as Roessink et al. (2013) but instead of summer generations, winter generations were tested. The 28d EC10 value (immobilisation) was 0.40 µg/L. The values for the winter generations are higher than for the summer generations. Therefore, the DS proposed to use for classification the lowest toxicity values for the summer generations, as most relevant for hazard assessment.

In one limit test with green alga Scenedesmus subspicatus, no effects were seen up to and including the highest dose tested, 10 mg/L. However no analytical monitoring was performed, so the effect value was based on nominal concentrations. In another limit study with Selenastrum capricornutum, the limit dose of 100 mg/L did have a statistically significant effect on growth rate, but this effect was < 50%. Therefore, NOErC value was <100 mg/L. Furthermore, the toxicity studies on the metabolite imidacloprid desnitro to Hyalella azteca (OECD TG 202) and to Chironomus riparius (OECD TG 219) showed that it is less toxic compared to parent compound.

Based on the 28d EC10 (immobilisation) of 0.00024 mg/L for Caenis horaria, the DS proposed classification as Aquatic Chronic 1 (M=1000).

In summary, the DS considered Imidacloprid as not rapidly degradable but not potentially bioaccumulative for classification purposes. The selected acute toxicity EC50 value is between 0.001-0.01 mg/L resulting a classification of Aquatic Acute 1 with M-factors of 100. The selected chronic toxicity EC10 value is between 0.00001-0.0001 mg/L, resulting a classification of Aquatic Chronic 1 and M=1000.

Comments received during public consultation

Four MSCAs commented the proposal and expressed a general agreement with the proposed classification based on mayflies. However, a commenting Company-Manufacturer raised several doubts on the non-standard species used to classify the substance. Most of the MSCAs also highlighted the need for more detailed information on the non-guideline key studies in order to verify the validity criteria and the actual concentrations of imidacloprid used in the tests.

The well-argued responses by the DS are reported in the “additional key elements” section in the Background Document.

Assessment and comparison with the classification criteria

Degradation RAC agrees with the DS proposal to consider imidacloprid as not rapidly degradable. The substance is hydrolytically stable and not ultimately degraded to a level greater than 70% over 28 days in surface water, water/sediment and soil simulation studies.

Bioaccumulation As experimentally determined BCF values are not available for imidacloprid, the assessment of bioaccumulation is based on experimentally determined log KOW value. Hence, based on the value of log KOW = 0.57 being below the decisive CLP criterion (log KOW < 4), RAC agrees with the DS proposal to consider the bioaccumulation potential of imidacloprid as low.

Aquatic toxicity Invertebrates are the most sensitive trophic level. The key study is non guideline and performed with non standard species, i.e. different species of mayflies. Insects have to be considered a representative group for the invertebrate trophic level, as the mode of action of imidacloprid implies acting as antagonist of the nicotinic acetylcholine receptor in the central nervous system of insects, thus disturbing synaptic signal transmissions of insects as Mayflies. Consequently, RAC considers the study relevant as well as reliable for use in classification.

Acute aquatic hazard

Acute aquatic toxicity studies are available for fish, invertebrates and algae. The most sensitive species were Cloeon dipterum and Caenis horaria in the same range of sensitivity. The key study is performed with non-standard different species of mayflies. Nevertheless, RAC considers the study relevant and reliable for use in classification. RAC concludes that, in order to provide consistency with the results from OECD TG 202, the 48h results can be used for classification. This does not change the proposal from that of the DS.

In conclusion, the most sensitive effect value is for Cloen dipterum (immobilisation). With a 48h

EC50 = 0.0027 mg/L, imidacloprid meets the classification as Aquatic Acute 1, M-factor=100, because the acute toxicity value is in the range 0.001 < EC50 ≤ 0.01 mg/L.

Chronic aquatic hazard Adequate chronic toxicity data is available for all three trophic levels. As for the acute toxicity, invertebrates are the most sensitive group. The lowest value is for mayflies with a 21d EC10 = 0.000056 mg/L (immobilisation). The key study is performed with a different non-standard species, nevertheless RAC considers the study relevant and reliable for use in classification. Similarly to the aquatic acute classification, RAC concludes that, in order to provide consistency with the results from OECD TG 211, the available 21d results can be used for classification. This does not change the proposal from that of the DS.

Imidacloprid fulfils the criteria for classification as Aquatic Chronic 1, M-factor =1000, because the chronic toxicity value is in the range of 0.00001 < NOEC ≤ 0.0001 mg/L and is considered not rapidly degradable.

In conclusion, RAC agrees with the DS that imidacloprid warrants classification as Aquatic Acute 1 (M=100) and Aquatic Chronic 1 (M=1000).

ANNEXES: Annex 1 The Background Document (BD) gives the detailed scientific grounds for the opinion. The BD is based on the CLH report prepared by the Dossier Submitter; the evaluation performed by RAC is contained in ‘RAC boxes’.

Annex 2 Comments received on the CLH report, response to comments provided by the Dossier Submitter and RAC (excluding confidential information).